Trikentramides A–D, Indole Alkaloids from the Australian Sponge

Article pubs.acs.org/jnp

Trikentramides A−D, Indole Alkaloids from the Australian Sponge Trikentrion flabelliforme Shahan Khokhar,† Yunjiang Feng,† Marc R. Campitelli,† Ronald J. Quinn,† John N. A. Hooper,†,‡ Merrick G. Ekins,‡ and Rohan A. Davis*,† †

Eskitis Institute, Griffith University, Brisbane, QLD 4111, Australia Queensland Museum, South Brisbane, QLD 4101, Australia

‡

S Supporting Information *

ABSTRACT: Chemical investigations of two specimens of Trikentrion f labelliforme collected from Australian waters have resulted in the identification of four new indole alkaloids, trikentramides A−D (9− 12). The planar chemical structures for 9−12 were established following analysis of 1D/2D NMR and MS data. The relative configurations for 9−12 were determined following the comparison of 1 H NMR data with data previously reported for related natural products. The application of a quantum mechanical modeling method, density functional theory, confirmed the relative configurations and also validated the downfield carbon chemical shift observed for one of the quaternary carbons (C-5a) in the cyclopenta[g]indole series. The indole-2,3-dione motif present in trikentramides A−C is rare in nature, and this is the first report of these oxidized indole derivatives from a marine sponge.

I

warranted further chemical investigation due to the presence of potentially new indole metabolites. Herein we report the isolation, structure elucidation, and density functional theory (DFT) NMR studies of four new indole alkaloids, namely, trikentramides A−D (9−12).

ndole alkaloids are ubiquitous natural products that have been shown to be produced by a variety of organisms, including marine invertebrates such as sponges.1,2 Over onequarter of marine alkaloids isolated to date contain the indole motif, and, in addition to displaying anticancer, antiprotozoal, antibacterial, antiviral, and anti-inflammatory activity, indole alkaloids often have unusual functionalities.1 Sponges belonging to the genus Trikentrion have been the source of several novel indole alkaloids. These include the antibacterial trikentrins 1−5 from the Australian T. f labelliforme,3 as well as trikendiol4 and trikentramine5 from the West African sponge T. leave. Although the trikentrins were the first report of cyclopenta[g]indoles from the marine environment, the related compounds herbindoles A−C (6−8) from an Axinella sp. have subsequently been reported.6 Previously isolated Trikentrion alkaloids displayed characteristic 1H NMR signals including downfield exchangeable indole protons, aromatic methine resonances, and in some cases upfield methyl doublet signals that corresponded to cyclopentyl-substituted methyl groups.3,6 This knowledge enabled us to rapidly dereplicate Trikentrion sponge extracts for the presence of known or new indole alkaloids. Using a modified solid-phase extraction (SPE) protocol developed in our research group, we were able to conduct NMR-based screening of semipurified Trikentrion extracts that enabled us to identify the presence of the indole chemotype.7,8 This protocol involved the rapid small-scale MeOH extraction of ground and freezedried sponge material followed by the removal of salts and highly lipophilic material using C18 SPE chromatography. Using this methodology we identified two T. f labelliforme samples that © 2013 American Chemical Society and American Society of Pharmacognosy

■

RESULTS AND DISCUSSION

Two ground and freeze-dried T. flabelliforme samples (300 mg) were subjected to rapid MeOH extraction followed by the removal of the salts and highly lipophilic material using C18 SPE chromatography (MeOH/H2O/1.0% TFA). The 1H NMR spectra in DMSO-d6 of both semipurified Trikentrion extracts displayed aromatic signals between δH 6.59 and 6.87 and, importantly, exchangeable NH proton signals between δH 10.18 and 11.12. Previous studies of indole alkaloids by our group have revealed that the exchangeable NH proton typically manifests between δH 10.0 and 12.0 in deuterated DMSO.9−12 HMBC experiments on the semipurified extracts revealed that the exchangeable and aromatic protons both experienced couplings to carbon atoms resonating downfield of δC 112 that were characteristic of sp2-hybridized carbon atoms, possibly suggestive of the indole chemotype. LC-MS analysis of the semipurified extracts revealed ions that did not correspond to any of the known Trikentrion alkaloids, which prompted further investigations.3−5 Received: July 29, 2013 Published: November 4, 2013 2100

dx.doi.org/10.1021/np400617h | J. Nat. Prod. 2013, 76, 2100−2105

Journal of Natural Products

Article

Table 2. 13C NMR Data for Trikentramides A−D (9−12) in DMSO-d6 δC, type position 2 3 3a 4 5 5a 6 7 8 8a 8b 9 10 11 12 a

A large-scale MeOH extraction of T. f labelliforme (voucher specimen: G303344) followed by C18 flash column chromatography (MeOH/H2O/1.0% TFA) and C18 HPLC (MeOH/ H2O/1.0% TFA) resulted in the isolation of two new metabolites, which we have named trikentramides A and B (9 and 10). Chemical investigations of an additional specimen of T. f labelliforme (voucher specimen G306125) using similar chromatographic methods resulted in the identification of two additional alkaloids, trikentramides C (11) and D (12). Trikentramide A (9, 0.4 mg, 0.012% dry wt) was isolated as an optically active yellow gum. The HRESIMS and 1D NMR data (Tables 1 and 2) enabled the establishment of a molecular

9a

10b

160.3, C 183.7, C 114.0, C 145.1, C 117.7, CH 161.0, C 39.1, CH 42.5, CH2 35.3, CH 128.5, C 146.4, C 24.4, CH2 14.5, CH3 20.7, CH3 20.8, CH3

11b

12a

c

c

183.6, C 113.9, C 145.7, C 117.7, CH 160.2, C 37.6, CH 42.4, CH2 34.0, CH 129.0, C 146.0, C 24.2, CH2 14.4, CH3 18.6, CH3 19.5, CH3

183.6, C 114.3, C 138.5, C 119.5, CH 160.5, C 38.6, CH 42.2, CH2 34.9, CH 128.1, C 146.0, C 17.3, CH3 20.4, CH3 20.4, CH3

177.4, C 34.5, CH2 122.5, C 131.4, C 117.1, CH 148.5, C 38.0, CH 43.4, CH2 35.8, CH 126.0, C 138.4, C 18.4, CH3 21.6, CH3 21.0, CH3

225 MHz at 25 °C. b150 MHz at 30 °C. cSignal not observed.

two carbonyl signals at δC 183.7 and 160.3. The HSQC correlation data enabled 1H NMR shifts to be assigned to each of their attached 13C atoms. Analysis of the COSY spectrum established two spin systems: a 1,3-dimethyl propyl group (δH 1.26, 3.06, 1.25, 2.54, 3.14, and 1.27) and an ethyl moiety (δH 1.13 and 2.83) (Figure 1). Weak COSY correlations from the aromatic proton

Table 1. 1H NMR Data for Trikentramides A−D (9−12) in DMSO-d6 δH, mult. (J in Hz) position NH 3 5 6 7a 7b 8 9 10 11 12 a

9a

10b

11b

10.83, s

11.03, s

10.83, s

6.77, s 3.06, ddq (7.2, 7.2, 7.2) 2.54, ddd (12.6, 7.2, 7.2) 1.25, ddd (12.6, 7.2, 7.2) 3.14, ddq (7.2, 7.2, 7.2) 2.83, q (7.6) 1.13, t (7.6) 1.26, d (7.2) 1.27, d (7.2)

6.75, s 3.26, m

6.74, s 3.04, ddq (7.6, 7.6, 7.6) 2.53, ddd (12.9, 7.6, 7.3) 1.24, ddd (12.9, 7.6, 7.3) 3.14, ddq (7.3, 7.3, 7.3) 2.42, s 1.26, d (7.6) 1.27, d (7.3)

1.95, ddd (12.9, 7.3, 1.9) 1.79, ddd (12.9, 9.0, 9.0) 3.21, m 2.83, 1.13, 1.23, 1.12,

q (7.6) t (7.6) d (7.2) d (7.2)

Figure 1. Key COSY, HMBC, and ROESY correlations for 9. 12a 10.18, s 3.31, s 6.59, s 2.99, ddq (7.2, 7.2, 7.2) 2.49, ddd (12.5, 7.3, 7.2) 1.15, ddd (12.5, 7.3, 7.2) 3.10, ddq (7.3, 7.3, 7.3) 2.14, s 1.22, d (7.2) 1.25, d (7.3)

(δH 6.77) to the methine (δH 3.06) and methylene (δH 2.83) in the two spin systems indicated that the propyl and ethyl moieties were attached to an aromatic ring. HMBC correlations confirmed this linkage, with strong cross-peaks (suggestive of a three-bond coupling in an aromatic system) observed from the aromatic proton at δH 6.77 (H-5) to the methylene carbon at δC 24.4 (C-9) and the methine carbon at δC 39.1 (C-6). Additional HMBC correlations from the two methine protons (δH 3.06 and 3.14) to the aromatic carbons [δC 161.0 (C-5a) and 128.5 (C-8a)] suggested the formation of a five-membered ring between the 1,3-dimethylpropyl unit and the benzenoid ring. The exchangeable proton at δH 10.83 showed an HMBC coupling to a carbonyl group at δC 183.7 (C-3), indicating the presence of a ketone functionality and also showed HMBC correlations to carbons resonating at δC 114.0 (C-3a) and 146.4 (C-8b) in the benzenoid ring. The last remaining and unassigned 13C signal at δC 160.3 suggested the presence of an additional carbonyl moiety in trikentramide A; however no HMBC correlations to this carbon were observed. Nonetheless, with all the atoms of 9 accounted for and the requirement to satisfy eight degrees of unsaturation, a pyrrolidine-2,3-dione fused to the benzenoid system was established. HMBC data (vide supra) secured the regioisomer assignment, which was further supported by a strong ROESY correlation between the

900 MHz at 25 °C. b600 MHz at 30 °C.

formula of C15H17NO2, with eight degrees of unsaturation. The 1 H NMR spectrum contained an upfield methyl triplet (δH 1.13), two methyl doublets (δH 1.26 and 1.27), two diastereotopic methylene protons (δH 2.54 and 1.25), two methine doublets of quartets (δH 3.14 and 3.06), a downfield aromatic methine singlet (δH 6.77), and a broad exchangeable proton (δH 10.83). The 13C NMR spectrum contained 15 carbon signals, including seven sp3 carbons, six sp2 carbons, and 2101

dx.doi.org/10.1021/np400617h | J. Nat. Prod. 2013, 76, 2100−2105

Journal of Natural Products

Article

NH proton at δH 10.83 and the H-12 protons at δH 1.27. Furthermore, comparison of the pyrrolidine-2,3-dione 13C chemical shifts in DMSO-d6 of 9 [δC 160.3 (C-2), 183.7 (C3), 114.0 (C-3a), 146.4 (C-8b)] with the commercially available natural product isatin (13) [δC 159.3 (C-2), 184.3 (C-3), 117.7 (C-3a), 150.7 (C-7a)]13 showed only minor differences. Thus the planar structure of trikentramide A was assigned as 9. Trikentramide B (10) was isolated as a 1:3 mixture (0.5 mg, 0.013% dry wt) with trikentramide A (9), as determined by 1H NMR analysis and integration of the aromatic methine protons [δH 6.77 (9) and 6.75 (10)]. Attempts to purify the mixture were not undertaken due to paucity of the sample. LC-MS analysis of the mixture revealed the presence of two compounds with identical molecular weights. The structure of 10 was determined via 1D and 2D NMR data analysis of the mixture. The 1H NMR data (Table 1) of 10 were very similar to those of 9; however major differences were noted in the chemical shifts of the diastereotopic (H-7a/H-7b) protons, which resonated at δH 1.95 and 1.79 in 10 compared to δH 2.54 and 1.25 in 9. These data suggested that 10 was a diastereomer of 9. The HRESIMS data for trikentramide C (11, 0.2 mg, 0.002% dry wt) gave a pseudomolecular ion at m/z 252.0997 [M + Na]+ that corresponded to a molecular formula of C14H14NO2 and eight degrees of unsaturation. This formula indicated that, in comparison to trikentramide A (9), 11 had one less methylene group in the molecule. Comparison of the 1H NMR data indicated that the ethyl group in 9 had been replaced by a methyl singlet in 11. The assignment was confirmed by the HMBC correlations from the methyl protons at δH 2.42 (H-9) to the carbon atoms resonating at δC 114.3 (C-3a), 138.5 (C4), and 119.5 (C-5). Therefore the planar structure of trikentramide C (11) was established. The final molecule, trikentramide D (12, 0.4 mg, 0.004% dry wt), gave a pseudomolecular ion at m/z 238.1205 [M + Na]+ in the HRESIMS measurement. This, in conjunction with 1H and 2D NMR data, established a molecular formula of C14H16NO with seven degrees of unsaturation. The loss of an oxygen atom and one degree of unsaturation for 12 compared to 11 suggested the absence of a carbonyl functionality at either C-2 or C-3. This was supported by the NMR data for 12 that showed only one carbonyl signal (δC 177.4) and an extra methylene resonance (δH 3.31/ δC 34.5). HMBC correlations were observed from the methylene protons (δH 3.31) to the amide carbonyl (δC 177.4, C-2) as well as the aromatic carbons C-3a (δC 122.5), C-4 (δC 131.4), and C-8b (δC 146.0). These data enabled the planar structure of trikentramide D to be assigned as 12. The relative configurations for 9−12 were established by comparison of their 1H NMR data to the known trikentrin analogues.3,6 Previous studies on the trikentrins (1−5) had shown that large 1H NMR chemical shift differences between the diastereotopic H-7 methylene protons (ΔδH‑7a/H‑7b > 1) were indicative of the cyclopentenyl unit containing methyl groups with a cis orientation, whereas no or minimal differences (ΔδH‑7a/H‑7b < 0.5) were diagnostic for the trans isomer.3,6,14 Trikentramides A, C, and D all had relatively large ΔδH‑7a/H‑7b (i.e., 1.29, 1.29, and 1.34, respectively), indicating that they all possessed the cis configuration at the C-6 and C-8 positions. Trikentramide B was assigned as having a trans configuration due to its small ΔδH‑7a/H‑7b value of 0.16. Of particular interest for this class of compounds was the extreme downfield chemical shift of C-5a (∼δC 160) in 9−11.

We wanted to validate this electron-deficient position and its corresponding deshielded 13C resonance by generating theoretical NMR chemical shifts. Density functional theory is a branch of physics that can be used to elucidate important properties concerning molecular structure.15 The generation of a theoretical electron density through DFT can be used to predict NMR chemical shifts.16 In some recent examples, this has confirmed regiochemistry of functional groups, as well as bond connectivity between atoms in a variety of small organic molecules.17,18 To calculate theoretical NMR shielding constants for the trikentramides, the corresponding diastereomeric structures were subject to molecular mechanics energy minimization and subsequent conformational searches using MMFFs.19 Each minimum energy conformer was further optimized using DFT with the B3LYP/6-31G(d,p) functional and basis set combination.20−23 Single-point calculations using the MPW1PW91/6-311+G(d,p) level of theory incorporating PBF implicit solvation24,25 were used to calculate NMR shielding constants via gauge including atomic orbitals (GIAO) methodology.26 A set of reference compounds with published experimental NMR data27 were used to generate a linear regression model correlating observed chemical shifts with the calculated shielding constants. The multireference regression model was subsequently used to scale the calculated shielding constants to provide 1H and 13C chemical shifts for both the cis and trans isomers of 9−12. The reported theoretical chemical shifts are the Boltzmann-weighted average of all values obtained for the conformers identified for each molecule. Analysis of the theoretical 13C NMR shifts of 9 generated by GIAO calculations enabled us to confirm the extreme downfield position of C-5a (Table 3). Using identical GIAO Table 3. Experimental and Theoretical (DFT-NMR) 13C NMR Data for 9a position

δC calcd

δC exptl

2 3 3a 4 5 5a 6 7 8 8a 8b 9 10 11 12 CMADb MDc

156.9 180.8 112.1 147.0 116.3 162.1 42.4 43.7 38.7 125.7 142.3 28.7 14.5 18.3 19.2 2.4 4.3

160.3 183.7 114.0 145.1 117.7 161.0 39.1 42.5 35.3 128.5 146.4 24.4 14.5 20.7 20.8

a

Experimental and theoretical data obtained in DMSO-d6. bCMAD = corrected mean absolute deviation. cMD = maximum deviation.

calculations (to those used for 9), the theoretical 13C NMR chemical shifts were also calculated for 10−12 (Supporting Information). The calculated values for C-5a in 10−12 were in good agreement with the assigned experimental values. Furthermore, an image displaying the relative electron density for 9 was generated by mapping the lowest unoccupied 2102

dx.doi.org/10.1021/np400617h | J. Nat. Prod. 2013, 76, 2100−2105

Journal of Natural Products

Article

molecules.3,14 The theoretical NMR data for the cis isomer of 11 were most consistent with the experimental data of 11, and again, the cis theoretical NMR data of 12 correlated with the experimental data of 12 (Supporting Information).28 The cyclopenta[g]indole structure class has rarely been reported, and the natural occurring trikentrins and herbindoles A−C are the only examples to date.32 While a substructure search of the Dictionary of Natural Products database using the indole-2,3-dione fragment identified 20 natural products from a variety of sources including bacteria, marine mollusks, and plants, 9−12 are the first report of these oxidized indole derivatives from a marine sponge.32 In addition the cyclopenta[g]indole-2,3-dione motif present in trikentramides A−C has never been reported.

molecular orbitals (LUMO) onto the electron density surface as seen in Figure 2. The image clearly indicated a relatively low electron density at C-5a compared with other aromatic carbons, hence explaining the downfield C-5a shift.

■

CONCLUSIONS A review of the literature together with the application of a simple and rapid SPE protocol coupled to NMR analysis has led to the identification of four new indole alkaloids from two specimens of Trikentrion f labelliforme. In addition to standard NMR and MS techniques, theoretical NMR calculations generated through DFT supported the structures.

Figure 2. The lowest unoccupied molecular orbital (LUMO) mapped onto the electron density surface for trikentramide A (9). Blue and orange coloring indicates electron deficiency.

■

There has been success in establishing and confirming relative configurations using DFT-GIAO NMR calculations coupled with the DP4 analysis in a number of cases.28−30 For example, the relative configuration of the plant alkaloid nobilisitine A was recently revised based on comparison of experimental NMR data and DFT-GIAO-generated NMR data, together with DP4 probability.29,31 Having already generated DFT data, it was simple to use the theoretical 1H NMR data to confirm our assignment of the relative configurations for 9−12. The DP4 analysis showed that the 1H NMR experimental data for 9 (Table 4) were most consistent with the theoretical Table 4. Experimental and Theoretical (DFT-NMR) 1H NMR Data for 9 and 10a δH exptl position

9

NH 5 6 7a 7b 8 9 10 11 12 CMADb MDc

10.83 6.77 3.06 2.54 1.25 3.14 2.83 1.13 1.26 1.27

δH calcd

δH calcd

cis

trans

d

d

6.61 3.02 2.27 1.16 3.09 2.81 1.01 1.12 1.12 0.12 0.27

6.59 3.22 1.78 1.79 3.07 2.70 1.01 1.09 1.00 0.11 0.17

EXPERIMENTAL SECTION

General Experimental Procedures. Optical rotations were recorded on a JASCO P-1020 polarimeter. UV spectra were recorded on a JASCO V-650 UV/vis spectrophotometer. NMR spectra were recorded at 30 °C on a 600 MHz Varian Unity INOVA spectrometer or at 25 °C on a 900 MHz Bruker NMR spectrometer. Both spectrometers were equipped with cryoprobes. The 1H and 13C chemical shifts were referenced to the solvent peaks for DMSO-d6 at δH 2.50 and δC 39.5, respectively. LRESIMS data were recorded on a Waters LCMS system equipped with a Luna C18 column (3 μm, 100 Å, 50 × 4.6 mm), a PDA detector, and a ZQ ESI mass spectrometer. HRESIMS were recorded on a Bruker MicrOTof-Q spectrometer (Dionex UltiMate 3000 Micro LC system, ESI mode). All solvents used for chromatography, [α]D, UV, and MS were Lab-Scan HPLC grade, and the H2O was Millipore Milli-Q PF filtered. A Waters 600 pump equipped with a Waters 966 PDA detector and Gilson 715 liquid handler were used for all HPLC work. Semipreparative HPLC was carried out with a Luna C18 column (5 μm, 100 Å, 250 × 10 mm). Alltech sample preparative C18-bonded silica (35−75 μm, 150 Å) was used for solid-phase extraction and flash column (100 × 30 mm) chromatography work. Phenomenex polypropylene cartridges (single frit, 3 mL) were used for the SPE work. Isatin was purchased from Sigma-Aldrich. Animal Material. Trikentrion f labelliforme (G303344) was collected at a depth of 10 m by scuba diving from East Point Bommies in the Northern Territory, Australia, in September 1993. The voucher specimen (G303344) was deposited at the Queensland Museum, South Brisbane, Queensland, Australia. T. f labelliforme (G306125) was collected at a depth of 28 m by trawl from approximately 70 km off the coast of Port Hedland, Western Australia, in September 1995, with the voucher specimen (G306125) also deposited at the Queensland Museum. This species and the genus were recently redescribed and revised.33 Both samples were immediately frozen upon collection and transported to the Eskitis Institute, where they were freeze-dried, ground, and stored as dried powders in airtight containers in the dark in a temperature-controlled environment (∼20 °C). Generation of Semipurified Sponge Extracts for NMR Screening. The freeze-dried and ground T. flabelliforme specimens (300 mg) were each added to two separate SPE cartridges, and the sponge material was flushed with MeOH (10 mL). The extracts were dried, preadsorbed onto C18-bonded silica (300 mg), and loaded onto a bed of C18-bonded silica (1.0 g) that was pre-equilibrated with MeOH/H2O/TFA (10:90:1). The column was flushed with MeOH/

δH exptl 10 11.03 6.75 3.26 1.95 1.79 3.21 2.83 1.13 1.23 1.12

a

Experimental and theoretical data obtained in DMSO-d6. bCMAD = corrected mean absolute deviation. cMD = maximum deviation. d Exchangeable signal not calculated.

chemical shifts calculated for the cis isomer, with a probability of greater than 99.9%. In contrast, the DP4 probability indicated that the experimental data for 10 were most consistent with the theoretical data calculated for the trans isomer (>99.9% probability). These assignments were consistent with results from previous studies of related 2103

dx.doi.org/10.1021/np400617h | J. Nat. Prod. 2013, 76, 2100−2105

Journal of Natural Products

Article

theoretical shifts reported are the Boltzmann-weighted average of values calculated for each conformer using the regression model.

H2O/TFA (10:90:1, 10 mL), which was discarded, followed by MeOH/H2O/TFA (70:30:1, 10 mL). The 70:30:1 MeOH/H2O/TFA fractions were dried and analyzed by 1H NMR spectroscopy. Extraction and Isolation. For the large-scale extraction of T. f labelliforme (G303344), a sample (3.3 g) of the freeze-dried and ground sponge was exhaustively extracted with MeOH (360 mL), and the resulting extract was dried via rotary evaporation to yield an orange-brown gum (0.5 g). This gum was preadsorbed onto C18bonded silica (3.0 g), placed on a bed of pre-equilibrated C18-bonded silica [9 g, MeOH/H2O/TFA (10:90:1)], and flushed with MeOH/ H2O/TFA (10:90:1, 200 mL), which was discarded, followed by MeOH/H2O/TFA (70:30:1, 200 mL). The latter eluent was dried (0.39 g), dissolved in MeOH (3 mL), and further purified by C18bonded silica semipreparative HPLC at a flow rate of 4 mL/min eluting with MeOH/H2O/TFA (75:25:1) for 5 min, followed by a linear gradient to MeOH within 30 min. Sixty fractions (0.5 min) were collected from the start of the HPLC run. Fractions 29 and 30/31 contained trikentramide A (9, 0.4 mg, 0.012% dry wt) and trikentramides A/B (9/10, 3:1 mixture, 0.5 mg, 0.013% dry wt), respectively. T. f labelliforme (G306125, 10 g) was also subjected to the same extraction conditions as G303344, yielding 0.4 g of extract. This was preadsorbed onto C18-bonded silica overnight and subjected to the same C18 column chromatography as G303344, yielding 0.14 g from the MeOH/H2O/TFA (70:30:1, 200 mL) wash. This was subjected to the same C18-bonded silica semipreparative HPLC conditions described for G303344. Fractions 20−21 and fractions 24−26 yielded trikentramides C (11, 0.2 mg, 0.002% dry wt) and D (12, 0.4 mg, 0.004% dry wt), respectively. Trikentramide A (9): yellow gum; [α]25D +40.7 (c 0.027, CHCl3); UV (MeOH) (log ε) 202 (3.80), 223 (3.56), 249 (3.41), 334 (2.95) nm; 1H and 13C NMR data (DMSO-d6) see Tables 1 and 2; (+)-LRESIMS m/z 266 [M + Na]+; (−)-LRESIMS m/z 242 [M − H] − ; (+)-HRESIMS m/z 266.1160 [M + Na] + (calcd for C15H17NO2Na, 266.1151). Trikentramide B (10): yellow gum (3:1 mixture of trikentramides A and B); 1H and 13C NMR data (DMSO-d6) see Tables 1 and 2.; (+)-LRESIMS m/z 266 [M + Na]+; (−)-LRESIMS m/z 242 [M − H]−; (−)-HRESIMS m/z 242.1191 [M − H]− (calcd for C15H16NO2, 242.1187). Trikentramide C (11): yellow gum; [α]25D +42.2 (c 0.028, CHCl3); UV (MeOH) (log ε) 202 (4.16), 223 (3.85), 250 (3.76), 333 (3.31) nm; 1H and 13C NMR data (DMSO-d6), see Tables 1 and 2; (+)-LRESIMS m/z 230 [M + H]+; (−)-LRESIMS m/z 228 [M − H] − ; (+)-HRESIMS m/z 252.0997 [M + Na] + (calcd for C14H15NO2Na, 252.0995). Trikentramide D (12): off-white powder; [α]25D +50.0 (c 0.02, CHCl3); UV (MeOH) (log ε) 218 (3.84), 255 (3.32), 333 (2.46) nm; 1 H and 13C NMR data (DMSO-d6), see Tables 1 and 2; (+)-LRESIMS m/z 216 [M + H]+; (+)-HRESIMS m/z 238.1205 [M + Na]+ (calcd for C14H17NONa, 238.1202). Computational Chemistry. All structures were subject to geometry optimization in MacroModel34 using the MMFFs with H2O solvation, with default parameters and convergence criteria used unless otherwise stated. A conformational search was performed on each structure using the mixed MCMM/low-mode search with 1000 steps per rotatable bond and an energy window of 21 kJ mol−1 for retention of conformers. All conformers identified by MacroModel were subject to further gas-phase geometry optimization in Jaguar35 using the B3LYP functional and 6-31G(d,p) basis set. A fine integration grid with accurate SCF convergence criteria was employed, and all geometry-optimized structures were characterized as true minima by the absence of imaginary vibrational frequencies at the stationary point. NMR shielding constants were calculated using the GIAO method as implemented in Jaguar using single-point MPW1PW91/6-311+G(d,p) calculations with the PBF implicit solvation model (DMSO). A multireference linear regression model was generated using the same computational methodology as described above for a range of commercially available compounds with experimental NMR shift data reported in DMSO-d6.27,36,37 The

■

ASSOCIATED CONTENT

S Supporting Information *

1D and 2D NMR spectra for trikentramides A−D, tabulated NMR data for isatin, scaling of computed chemical shifts, energies and Cartesian coordinates, tables of experimental and theoretical 1H and 13C NMR shifts, and DP4 results. This material is available free of charge via the Internet at http:// pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Tel: +61-7-3735-6043. Fax: +61-7-3735-6001. E-mail: r. davis@griffith.edu.au. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank G. MacFarlane from the University of Queensland for acquiring the HRESIMS measurements. We also would like to thank G. Pierens from the Centre for Advanced Imaging (University of Queensland) for acquiring the 900 MHz NMR data. S.K. would like to thank the Australian Government for an Australian Postgraduate Award (APA) scholarship.

■

REFERENCES

(1) Gul, W.; Hamann, M. T. Life Sci. 2005, 78, 442−453. (2) Blunt, J. W.; Copp, B. R.; Keyzers, R. A.; Munro, M. H. G.; Prinsep, M. R. Nat. Prod. Rep. 2012, 29, 144−222. (3) Capon, R. J.; Macleod, J. K.; Scammells, P. J. Tetrahedron 1986, 42, 6545−6550. (4) Loukaci, A.; Guyot, M. Tetrahedron Lett. 1994, 35, 6869−6872. (5) Aknin, M.; Miralles, J.; Kornprobst, J.-M.; Faure, R.; Gaydou, E.M.; Boury-Esnault, N.; Kato, Y.; Clardy, J. Tetrahedron Lett. 1990, 31, 2979−2982. (6) Herb, R.; Carroll, A. R.; Yoshida, W. Y.; Scheuer, P. J.; Paul, V. J. Tetrahedron 1990, 46, 3089−3092. (7) Camp, D.; Davis, R. A.; Campitelli, M.; Ebdon, J.; Quinn, R. J. J. Nat. Prod. 2012, 75, 72−81. (8) Wright, A. E. In Natural Products Isolation Methods in Biotechnology; Cannell, R. J. P., Ed.; Humana Press: NJ, 1998; Vol. 4, pp 365−408. (9) Feng, Y.; Davis, R. A.; Sykes, M. L.; Avery, V. M.; Quinn, R. J. Bioorg. Med. Chem. Lett. 2012, 22, 4873−4876. (10) Davis, R. A.; Duffy, S.; Avery, V. M.; Camp, D.; Hooper, J. N. A.; Quinn, R. J. Tetrahedron Lett. 2010, 51, 583−585. (11) Rudolph, K. E.; Liberio, M. S.; Davis, R. A.; Carroll, A. R. Org. Biomol. Chem. 2013, 11, 261−270. (12) Liberio, M. S.; Sooraj, D.; Williams, E. D.; Feng, Y.; Davis, R. A. Tetrahedron Lett. 2011, 52, 6729−6731. (13) Full NMR data of isatin can be obtained in the Supporting Information. (14) Ashraf, C.; Jackson, W. Aust. J. Chem. 1978, 31, 1845−1849. (15) Kohn, W.; Becke, A. D.; Parr, R. G. J. Phys. Chem. 1996, 100, 12974−12980. (16) Lodewyk, M. W.; Siebert, M. R.; Tantillo, D. J. Chem. Rev. 2011, 112, 1839−1862. (17) Bagno, A.; Rastrelli, F.; Saielli, G. Chem. Eur. J. 2006, 12, 5514− 5525. (18) Pu, J.-X.; Huang, S.-X.; Ren, J.; Xiao, W.-L.; Li, L.-M.; Li, R.-T.; Li, L.-B.; Liao, T.-G.; Lou, L.-G.; Zhu, H.-J.; Sun, H.-D. J. Nat. Prod. 2007, 70, 1706−1711. (19) Halgren, T. A. J. Comput. Chem. 1999, 20, 720−729. (20) Becke, A. D. J. Chem. Phys. 1993, 98, 5648−5652.

2104

dx.doi.org/10.1021/np400617h | J. Nat. Prod. 2013, 76, 2100−2105

Journal of Natural Products

Article

(21) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785−789. (22) Vosko, S. H.; Wilk, L.; Nusair, M. Can. J. Phys. 1980, 58, 1200− 1211. (23) Stephens, P. J.; Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J. J. Phys. Chem. 1994, 98, 11623−11627. (24) Tannor, D. J.; Marten, B.; Murphy, R.; Friesner, R. A.; Sitkoff, D.; Nicholls, A.; Honig, B.; Ringnalda, M.; Goddard, W. A. J. Am. Chem. Soc. 1994, 116, 11875−11882. (25) Marten, B.; Kim, K.; Cortis, C.; Friesner, R. A.; Murphy, R. B.; Ringnalda, M. N.; Sitkoff, D.; Honig, B. J. Phys. Chem. 1996, 100, 11775−11788. (26) Adamo, C.; Barone, V. J. Chem. Phys. 1998, 108, 664−675. (27) Gottlieb, H. E.; Kotlyar, V.; Nudelman, A. J. Org. Chem. 1997, 62, 7512−7515. (28) Smith, S. G.; Goodman, J. M. J. Am. Chem. Soc. 2010, 132, 12946−12959. (29) Lodewyk, M. W.; Tantillo, D. J. J. Nat. Prod. 2011, 74, 1339− 1343. (30) Wyche, T. P.; Hou, Y.; Braun, D.; Cohen, H. C.; Xiong, M. P.; Bugni, T. S. J. Org. Chem. 2011, 76, 6542−6547. (31) Schwartz, B. D.; Jones, M. T.; Banwell, M. G.; Cade, I. A. Org. Lett. 2010, 12, 5210−5213. (32) Dictionary of Natural Products on DVD, version 21.2; Taylor & Francis/CRC Press: London, UK, 2012. (33) Van Soest, R.; Carballo, J. L.; Hooper, J. ZooKeys 2012, 239, 1− 70. (34) MacroModel, version 9.9, Suite 2012; Schrödinger, LLC: New York, NY, 2012. (35) Jaguar, version 7.9, Suite 2012; Schrödinger, LLC: New York, NY, 2012. (36) Bally, T.; Rablen, P. R. J. Org. Chem. 2011, 76, 4818−4830. (37) Jain, R.; Bally, T.; Rablen, P. R. J. Org. Chem. 2009, 74, 4017− 4023.

2105

dx.doi.org/10.1021/np400617h | J. Nat. Prod. 2013, 76, 2100−2105